RIAA CEO Mitch Glazier argued in a November 9 interview on The Verge's Decoder podcast that music generated by Suno substitutes for human-created tracks, forming the core of the labels' opposition to the AI company's fair use defense.
๐ Why substitution is the key legal argument
Glazier told host Nilay Patel that the fourth fair use factor โ the effect on the potential market for the original โ is central to the RIAA's case against Suno. He explained that AI-generated songs compete on the same streaming platforms, right next to human music, for the same subscription fee. "If what you're creating substitutes for something else, it's probably not fair use," Glazier stated. This framing positions Suno's output as direct market competition rather than transformative use.
โ๏ธ How this fits the broader Suno lawsuits
The comments come as Sony Music, Universal Music Group and others continue litigation against Suno in Massachusetts federal court. Warner has settled and licensed with Suno, but the remaining majors argue that even v6 models built with some licensed data still raise infringement issues. Glazier described the case as precedent-setting, aimed at establishing that training on copyrighted music without permission is not fair use. He also noted the RIAA is pursuing a new federal voice-and-likeness right via the No Fakes Act and pushing for track-level labeling of AI content on platforms.
๐ The massive scale mismatch driving urgency
Glazier highlighted the economic disparity: the entire U.S. recorded music business generated roughly $11 billion in wholesale revenue last year, while a single AI company's market cap reaches $2 trillion. With the RIAA operating with just 46 employees, the interview underscores the challenge of policing AI at industrial scale. He suggested that labeling for fully AI-generated music should begin at distributors rather than relying solely on services like Spotify.
๐๏ธ What this means for your next song
As legal pressure mounts on unlicensed training data, Suno creators should prioritize originality and lean into tools built for licensed or user-driven workflows. Start every project by clearly defining a unique concept instead of artist clones, then use /genie to name a record and instantly generate a full song kit complete with style, voice, structure, chords and mastering notes. This keeps your output distinct, avoids guardrail triggers, and positions your tracks for better streaming performance amid growing industry scrutiny. Cross-reference generated lyrics against real emotional intent, experiment with v6's section editing for precise changes, and upload your own reference audio only when it is fully original. These steps future-proof your catalog as platforms tighten rules around AI content.
Bottom line: Glazier's substitution argument signals that courts and labels see AI music as direct competition, making official licensing partnerships the clearest path forward for sustainable AI song creation.